Wing attach bolts in tension

Homebuilt Aircraft & Kit Plane Forum

Help Support Homebuilt Aircraft & Kit Plane Forum:

mcrae0104

Well-Known Member
Supporting Member
Joined
Oct 27, 2009
Messages
4,284
Location
CO
References/examples?
See post #70 above.

Every engineering doc I've ever seen related to aircraft says that joint friction is ignored by the designer of structural shear joints.

The primary function of the bolt in a shear connection is to clamp the materials together. Then one should consider how the joint may fail if the bolt(s) are not preloaded properly or if preload is lost in service.

Shigley has been mentioned as another reference:

1660968160309.png
The few strut braced attachments I've seen use clevis pin type attachments; in some cases using a literal clevis pin & clip, with zero clamping force.

The idea of ignoring friction is not rooted in the idea that the bolt's primary purpose isn't to act as a clamp; rather, it is a conservative design practice that considers methods of failure that come into play without friction. Now once we assume there is no preload, the parts may slip cyclically (as seen in the wheel picture posted earlier) and the bearing surfaces (or the bolt) may fail. Specifying a close-tolerance connection is another layer of conservative design practice (belt, suspenders, and duct tape). Even with a close-tolerance bolt without preload, some of the bearing occurs on the threaded portion of the bolt--this is another reason that disregarding preload and friction is not ideal.

One may be able to get away with a close-tolerance joint without preload in a low-stress application, but it is not best practice for joints we depend on for human life.
 
Last edited:

Bellaire MK

Active Member
Joined
Oct 11, 2021
Messages
44
You are correct. All attach points are considered pivot points and engineered as such.
 

rv7charlie

Well-Known Member
Supporting Member
Joined
Nov 17, 2014
Messages
3,365
Location
Pocahontas MS
OK, I don't have any formal engineering training, and I haven't stayed in any hotels for over 2 years. And I really want to learn. But some of the above seems very inconsistent with what we see in aircraft fastening systems the real world. One glaring thing (at least to me) is:
Even with a close-tolerance bolt without preload, some of the bearing occurs on the threaded portion of the bolt
Maybe I'm having trouble parsing this; perhaps it's the added phrase 'without preload'. All my informal training has said that the threads shouldn't be within the shear-loaded area of the joint, so the threads *don't* take any of the shear loads. Then there's the example of strut braced wings using literal clevis pins at their joints. At least in the case of Van's designs, the spec to use a castellated nut on the rear spar/fuselage attach bolt, because of the possibility (expectation?) that the joint can rotate with wing loading/unloading; the same caution given for any joint that can be loaded in rotation (sorry for not having the correct engineering term for that).

I *know* (from direct, personal experience) the clamping idea is ineffective with, for instance, IVO props. Back in the 1990s (my last experience with them) IVO prop hubs had no drive lugs for the blades. The blades had loose-fitting holes for the prop bolts. When assembled to IVO's torque specs, the blades had a pretty strong tendency to move in the hub, to the point that IVO had to issue service bulletins about it, then they knurled the faces of the hub clamping plates, and *still* had blade movement when running on direct drive 4 cyl engines, and eventually quit selling the props for Lycs because they couldn't solve the blade movement issue. Now I realize that this isn't a steel-on-steel or steel/aluminum joint (it's aluminum/fiberglass interface), but is there anyone on this forum who would risk running a constant speed prop on a Lyc, without the drive lugs? Based on the clamping theory, it should work just fine.
 

mcrae0104

Well-Known Member
Supporting Member
Joined
Oct 27, 2009
Messages
4,284
Location
CO
All my informal training has said that the threads shouldn't be within the shear-loaded area of the joint, so the threads *don't* take any of the shear loads.
Yes, that's good practice, and the thread length on AN bolts is designed to accommodate this. A slightly longer bolt and a second washer under the nut can accomplish this. (This is convenient since two washers equals 1/8" and AN bolts come in 1/8" increments--see AC43-13 on stacked washers.) I only mentioned bearing on threads because many bolts have a longer thread length than AN hardware, resulting in either lots of extra thread outside the joint, or someone not knowing any better and leaving threads inside the bearing portion of the joint, which is common practice (at least outside of aviation) on joints not designed for shear.
At least in the case of Van's designs, the spec to use a castellated nut on the rear spar/fuselage attach bolt, because of the possibility (expectation?) that the joint can rotate with wing loading/unloading; the same caution given for any joint that can be loaded in rotation (sorry for not having the correct engineering term for that).
The castellated nut makes sense to safety the nut given the potential for rotation caused by cyclic load reversal; Van's probably also has included this to ensure not only that the nut stays in place, but so that preload is maintained. I shouldn't really armchair-engineer their design, but if I were designing a similar joint, I would make it preloaded and check for a very clean set of faying surfaces before assembly. (I bet there's a torque value called out on that bolt, and it wouldn't surprise me if they have either a prohibition on, or guidance provided on, faying-surface coatings.)

I'm not a mechanical engineer, but in the construction industry, that connection would be need to be considered at least pre-tensioned, or better yet, slip-critical (as opposed to merely snug-tight, disregarding friction). Here is a specification which, although not completely applicable to aerospace, sheds some light on the principles in play.

In buildings--and I imagine this carries over to mechanical engineering more generally--pretensioning is called for in a joint that is subject to a) significant load reversal or b) where the bolt may be subject to combined shear and tension (as may be the case if the joined plates deform--no bueno). A slip-critical joint (i.e. pretensioned with particular attention paid to the smoothness/cleanliness of the faying surfaces) is appropriate where slippage (in the case of the rear spar joint, rotation) would be detrimental to the faying surfaces or where fatigue could result from load reversals.

Best practice notwithstanding, experience does show us that some joints work where bolt tension is negligible or nonexistent, as you point out. Those joints are either a) not subject to the considerations above, or b) carry a significant factor of safety (lightly stressed), and c) hopefully readily accessible for inspection.

I *know* (from direct, personal experience) the clamping idea is ineffective with, for instance, IVO props. Back in the 1990s (my last experience with them) IVO prop hubs had no drive lugs for the blades. The blades had loose-fitting holes for the prop bolts. When assembled to IVO's torque specs, the blades had a pretty strong tendency to move in the hub, to the point that IVO had to issue service bulletins about it, then they knurled the faces of the hub clamping plates, and *still* had blade movement when running on direct drive 4 cyl engines, and eventually quit selling the props for Lycs because they couldn't solve the blade movement issue. Now I realize that this isn't a steel-on-steel or steel/aluminum joint (it's aluminum/fiberglass interface), but is there anyone on this forum who would risk running a constant speed prop on a Lyc, without the drive lugs? Based on the clamping theory, it should work just fine.

I'm sorry I can't speak to that as I've never dived into prop attachment design. The best I can do is speculate that in addition to providing increased bearing area, the drive lug is there to carry the shear before the tensioned bolt experiences it and becomes subject to shear and bending in addition to axial preload. I'm sure others can contribute more on this than I can.
 

pylon500

Well-Known Member
Joined
Dec 26, 2003
Messages
608
Location
Taree Airport Australia
Hi PC,

There are spherical rod ends on the studs. The good question or concern is the bending loads imposed when the aircraft is pulled or pushed on the struts when ground handling the aircraft. I have been working on a KF-7 and am aware of the design details. This is why Piper has increased the strut forks to 5/8dia. on the J-3 and all others.
Designing with bolts in shear is optimum for light aircraft. As for the North American AT-6 cantilever wing attach of the outer panels is at the outer most surface of the airfoil, a multitude of bolts in tension on the flanges.
We are currently designing and building a new and more conventional wing for the Aircam that will create an alternative to the factory wing. Here are some photos. Best regards, Rick Berstling
Discussion of this 'new' wing probably deserves another thread, but from what I can see, this 'new' wing isn't really achieving anything over the original wing...?
 

Hawk81A

Well-Known Member
Joined
Sep 3, 2021
Messages
342
Slight correction here on washer thicknesses... AN/NAS/MS series washers come in at least 5 different thicknesses, depending on size.
Gets better all the time. I was concerned when someone indicated it MIGHT be okay to have threads in the fitting (?). I don't like this, even on nonaviation applications. I was aware, although it's been a long time that AN hardware was available in 1/8 increments, that there were various thicknesses of washers, that you were limited to a certain number of washers (2), and a certain number of exposed threads (was it 3?). Dennis
 

cvairwerks

Well-Known Member
Joined
May 12, 2010
Messages
424
Location
North Texas
Gets better all the time. I was concerned when someone indicated it MIGHT be okay to have threads in the fitting (?). I don't like this, even on nonaviation applications. I was aware, although it's been a long time that AN hardware was available in 1/8 increments, that there were various thicknesses of washers, that you were limited to a certain number of washers (2), and a certain number of exposed threads (was it 3?). Dennis
Typical engineering standards are for a total of 3 washers on a single bolt or screw, but, there are cases where there are different callouts. MS spline bolts have a radius under the head and must be used with a special radiused washer, so that it sits flat on the surface. Often, this is then combined with up to three washers for the rest of the installation. There is also the case of a type of nut that has a concave faced washer that is used with it to sit flat on a surface, where a bolt or screw hole was not drilled perpendicular to the face. I've forgotten the designation on them, but they allow a bolt to have something like a 10 degree angle thru the structure. Another oddball case, is where the designer calls out a peel washer along with multiple regular washers to act as a gap filler on a non-moving installation. The peel washer is often up to .090" thick and is constructed of .003" laminations. It's peeled to provide the exact thickness required, when standard washer thickness cannot be combined to generate the needed dimension.

As a couple of outliers on the standard design, I have installations that I deal with, that use 5, 7, 9 or 16 washers in the hardware stack, with no substitutions allowed. The engineering criteria allows some flexibility on where specific thicknesses can go, but no other variations.
 

wsimpso1

Super Moderator
Staff member
Joined
Oct 18, 2003
Messages
9,476
Location
Saline Michigan
Since friction is the holding force in aviation shear joints, I wonder; what's the purpose of A/N 320 and A/N364 nuts?
Regular AN bolts and nuts are intended for tightening to a large fraction of shank yield. Regular bolts and nuts have enough material in bolt heads and nuts to carry those large loads that make for friction.

These nuts (AN320 and 364) are used to keep the bolt from falling out of the joint. They go with bolts with reduced height heads (NAS1103 and 1106) and clevis bolts (AN23 and 27) that are primarily loaded in shear. They are pins in joints with loads carried through the bolts in shear and where movement is anticipated and/or intended. Torques used on these fasteners are usually near zero, friction imposed zero, they will easily loosen a regular nut in use, so they all are either castellated or self-locking. Since large preload from torque is not part of their use, these fasteners can tolerate thinner heads and nuts, allowing weight save. Somebody else thinks that weight is the enemy too...

Billski
 

cvairwerks

Well-Known Member
Joined
May 12, 2010
Messages
424
Location
North Texas
The 320's and 364's are also used where there is low clamping for required and clearance issues. There are places where a standard height nut would either not be able to be installed or would protrude into an area that would cause interference with another installation. A good example for the 320, is on a bellcrank arm, where the end will pass close to structure, or some other moving part.
 

wsimpso1

Super Moderator
Staff member
Joined
Oct 18, 2003
Messages
9,476
Location
Saline Michigan
is loss of preload a result of high stress and deformation????? and that is why we ignore friction in our calculations?
Loss of preload comes from:
Inadequate assembly/tightening process;
Poor design/assembly specifications;
Inadequate knowledge of loads/vibe/thermal situation;
Inadequate fastener design knowledge;
Improper parts/material substitution,

Properly spec’ed, designed, and installed joints do not loosen or fail. If joints fail, somebody or multiple somebodies failed at their jobs.

Sometimes it happens that the designer suspects the preload of the joint will be compromised. Design for shear with bolts that are installed in match drilled holes and close tolerance bolts can be done. With nominally toleranced hole sizes and positions, only two bolts in any set can be counted upon to bear before yielding occurs, which means you are attempting to design in the yielding range, which also means significant fatigue exposure once that occurs. This is dangerous territory.

Good joint design is the best scheme, it can have redundancy, can be made with large FOS and still be reasonably light, and even be made tolerant of wide ranges of materials.

Next time, I will discuss how overtightening effects joints.

Billski
 

Bellaire MK

Active Member
Joined
Oct 11, 2021
Messages
44
We were removing wings from a J-3 and noticed the upper strut ends had been crushed and the strut attach channels were deformed from over torque of the bolts. The new sealed struts have a crush bushing in place to prevent this. Torque specs would apply provided there is a crush spacer or bushing.
 

proppastie

Well-Known Member
Supporting Member
Joined
Feb 19, 2012
Messages
5,801
Location
NJ
If joints fail, somebody or multiple somebodies failed at their jobs
With the way engineers like to calculate and test things I would have though there would be all kinds of charts as regard friction and joint strength.....what charts I have found point to traditional calculation ignoring friction.....perhaps if I factor in fatigue exposure the calculated numbers make more sense?
 
Top